To compare the outcome among patients with invasive bladder cancer treated with cystectomy alone with outcome among those treated with combined-modality treatment in a randomised phase III trial.

Patients with histologically confirmed invasive non-metastatic bladder cancer T2-3, N0 and M0 were randomly assigned to two arms: Arm 1: of which all patients underwent radical cystectomy (RC) alone; and Arm 2, of which all patients were subjected to maximal transurethral resection of bladder tumour, followed 2 weeks later by combined chemoradiotherapy. The whole pelvis received 46 Gy in 23 fractions over 4·5 weeks. Chemotherapy was administered concomitantly with radiotherapy with: cisplatin 70 mg/m2 q. 3 weeks and Gemcitabine 300 mg/m2 D 1, 8 and 15 q. 3 weeks for two cycles. Patients who had complete response were shifted to phase II treatment: 20 Gy/10 fractions/2 weeks to the bladder. Patients with residual tumour underwent RC.

Of the 80 patients assigned Arm 2, a visibly completed transurethral resection of the bladder tumour was possible in 48 patients (60%). Phase I of combined chemoradiotherapy (CCRT) was accomplished in 74 patients. Post-induction urologic evaluation revealed no evidence of disease in 62 patients (83·8%) and residual disease in 12 patients (16·2%). Phase II of CCRT was completed in 58 of the 62 patients. The median follow-up for all patients is 27 months (range: 4–49). The 3-year overall survival (OS) for the combined-modality group and for the surgery group were 61 and 63%, respectively (p = 0·425), whereas the disease-specific survival (DSS) for each group was 69 and 73%, respectively (p = 0·714). The 3-year OS with bladder preservation for Arm 2 patients was 50%.

Multivariate analysis for the whole series showed that tumour stage and performance status (PS) were the only factors independently associated with DSS, although PS was the only factor independently associated with OS. In addition, residual disease after transurethral resection of the bladder tumour in Arm 2 patients was independently associated with both DSS and OS.

Acute toxicity was moderate and most of the late toxicities were grade 2 with no grade 4 toxicity and no treatment-related deaths, none required cystectomy for bladder contraction.

This study demonstrates that trimodality bladder-preserving approach represents a valid alternative for suitable patients. The OS and DSS rates of patients treated with trimodality bladder-preserving protocol are comparable to the results reported on patients treated with immediate radical cystectomy.